stumpf@gtenmc.UUCP (Jon S. Stumpf) (02/22/90)
How do I calculate maximum effective cable length for a given baud rate?
Also, is there a function to provide error rates (eg. 2% error) for the
given table?
Length
10 20 30 40 50 60 70 80 90 100
Baud +--------------------------------------------------------------
|
300 |
|
1200 |
|
2400 |
|
9600 |
|
19200 |
|
38400 |
+--------------------------------------------------------------
--
jss - Jon S. Stumpfroy@phri.nyu.edu (Roy Smith) (02/23/90)
> a function to provide error rates (eg. 2% error) for the given table? > 10 20 30 40 50 60 70 80 90 100 > [...] > 9600 Assuming your lengths are in feet, you had better extend the table a lot to the right. We have runs of up to about 250 feet of Belden 8723 (shielded 24 (22?) gauge twisted pair) which we run at 9600 with, as far as I can tell, 0% error rate with no special line drivers. I suspect that if you are willing to put up with a 2% error rate (which sounds pretty drastic to me) you could probably get many hundreds of feet at 9600, if you use the right kind of wire. With line drivers, you should be able to get thousands of feet, but that's not RS-232 any more. -- Roy Smith, Public Health Research Institute 455 First Avenue, New York, NY 10016 roy@alanine.phri.nyu.edu -OR- {att,philabs,cmcl2,rutgers,hombre}!phri!roy "My karma ran over my dogma"
dmt@pegasus.ATT.COM (Dave Tutelman) (02/23/90)
In article <608@gtenmc.UUCP> stumpf@gtenmc.UUCP (Jon S. Stumpf) writes: > >Also, is there a function to provide error rates (eg. 2% error) for the >given table? > > Length > 10 20 30 40 50 60 70 80 90 100 > Baud +-------------------------------------------------------------- > | > 300 | > | > 1200 | >......... >19200 | > | >38400 | > +-------------------------------------------------------------- The standard is quite clear; up to 15 feet and 20 kilobaud, there should be zero errors. Beyond that, you're on your own. In order to meet the standard, any IMPLEMENTATION has to allow some safety margin. How much is up to the designer. Most existing commercial interface chips build in a lot of margin. But it is generally margin, not an advertised feature. So..... Experiment with your favorite impementation, then TAKE SOME DESIGN RESPONSIBILITY. Decide, on the basis of your tests, how much you want to cut the safety margin the designer built in. Seriously, unless the supplier of your implementation advertises specs other than "RS-232 interface", you can't get any more specific. +---------------------------------------------------------------+ | Dave Tutelman | | Physical - AT&T Bell Labs - Lincroft, NJ | | Logical - ...att!pegasus!dmt | | Audible - (201) 576 2194 | +---------------------------------------------------------------+